Magnetic resonance imaging-based tumour perfusion parameters are biomarkers predicting response after radiation to brain metastases

R Jakubovic; A Sahgal; H Soliman; R Milwid; L Zhang; A Eilaghi; R I Aviv

doi:10.1016/j.clon.2014.06.010

Magnetic resonance imaging-based tumour perfusion parameters are biomarkers predicting response after radiation to brain metastases

Clin Oncol (R Coll Radiol). 2014 Nov;26(11):704-12. doi: 10.1016/j.clon.2014.06.010. Epub 2014 Jul 10.

Authors

R Jakubovic¹, A Sahgal², H Soliman², R Milwid³, L Zhang³, A Eilaghi³, R I Aviv⁴

Affiliations

¹ Department of Medical Imaging, Division of Neuroradiology, Sunnybrook Health Sciences Centre, University of Toronto, Toronto, Ontario, Canada; Department of Biomedical Physics, Ryerson University, Toronto, Ontario, Canada.
² Department of Radiation Oncology, Odette Cancer Centre of the Sunnybrook Health Sciences Centre, University of Toronto, Toronto, Ontario, Canada.
³ Department of Medical Imaging, Division of Neuroradiology, Sunnybrook Health Sciences Centre, University of Toronto, Toronto, Ontario, Canada.
⁴ Department of Medical Imaging, Division of Neuroradiology, Sunnybrook Health Sciences Centre, University of Toronto, Toronto, Ontario, Canada. Electronic address: richard.aviv@sunnybrook.ca.

PMID: 25023291
DOI: 10.1016/j.clon.2014.06.010

Abstract

Purpose: To investigate whether early relative cerebral blood volume (rCBV), relative cerebral blood flow (rCBF) and permeability (Ktrans(2)) measurements may serve as magnetic resonance imaging (MRI) biomarkers of radiation response or progression for brain metastases.

Materials and methods: Seventy brain metastases in 44 patients treated with either stereotactic radiosurgery or whole brain radiotherapy were imaged with dynamic susceptibility and dynamic contrast enhancement MRI at baseline, 1 week and 1 month after treatment. The final response status was determined according to volume criteria derived from a 1 year post-treatment MRI or last available follow-up MRI. Tumours were characterised as responders, non-responders, progressors and non-progressors and compared for Ktrans(2), rCBF and rCBV differences. Uni- and multivariate analysis evaluated factors associated with tumour response and progression at 1 week and 1 month. A generalised estimating equations (GEE) model accounted for multiple tumours per subject. Receiver operator characteristic (ROC) analysis identified optimal cut-off values, sensitivity and specificity for response or progression.

Results: Tumour responders showed lower Ktrans(2) and reduced rCBF at 1 week (P < 0.05 each). Progressive disease showed lower rCBF and reduced rCBV at 1 month (P < 0.05 each). GEE and multivariate analysis revealed lower Ktrans(2) at 1 week, an absence of prior radiation predicted response. At 1 month only lower rCBV predicted progressive disease on GEE and multivariate analysis. Optimal cut-off points for Ktrans(2) and rCBV were 1.37 and 2.03 with sensitivity and specificity of 61.5 and 81.1% and 73.9 and 81.8%, respectively.

Conclusion: Lower Ktrans(2) at 1 week and rCBV at 1 month discriminated responders and progressive disease, respectively.

Keywords: Cancer; dynamic contrast-enhanced imaging; dynamic susceptibility imaging; magnetic resonance perfusion; magnetic resonance permeability; metastasis.

Publication types

Clinical Trial, Phase II
Observational Study

MeSH terms

Aged
Biomarkers / analysis*
Blood Volume / radiation effects*
Brain Neoplasms / blood supply
Brain Neoplasms / radiotherapy*
Brain Neoplasms / secondary
Brain Neoplasms / surgery
Cerebrovascular Circulation / radiation effects*
Disease Progression
Female
Follow-Up Studies
Humans
Magnetic Resonance Imaging / methods*
Male
Middle Aged
Neoplasm Staging
Perfusion
Prognosis
Prospective Studies
Radiation Injuries / diagnosis
Radiation Injuries / etiology
Radiosurgery / adverse effects*
Radiotherapy / adverse effects*

Substances

Biomarkers